RESEARCH ARTICLE
Infectiousness of Sylvatic and Synanthropic Small Rodents Implicates a Multi-host Reservoir of Leishmania (Viannia) braziliensis Maria S. Andrade1,2☯, Orin Courtenay3☯, Maria E. F. Brito1, Francisco G. Carvalho1, Ana Waléria S. Carvalho1, Fábia Soares1, Silvia M. Carvalho1, Pietra L. Costa1, Ricardo Zampieri4, Lucile M. Floeter-Winter4, Jeffrey J. Shaw5‡*, Sinval P. BrandãoFilho1‡* 1 Department of Immunology, Aggeu Magalhães Research Center-Oswaldo Cruz Foundation (FIOCRUZ), Recife, Pernambuco, Brazil, 2 University of Pernambuco, Recife, Pernambuco, Brazil, 3 Warwick Infectious Disease Epidemiology Research (WIDER) and School of Life Sciences, University Warwick, Coventry, United Kingdom, 4 Biosiences Institute, São Paulo University, São Paulo, São Paulo, Brazil, 5 Biomedical Institute, São Paulo University, São Paulo, São Paulo, Brazil ☯ These authors contributed equally to this work. ‡ JJS and SPBF also contributed equally to this work. * [email protected] (JJS); [email protected] SPBF OPEN ACCESS Citation: Andrade MS, Courtenay O, F. Brito ME, Carvalho FG, Carvalho AWS, Soares F, et al. (2015) Infectiousness of Sylvatic and Synanthropic Small Rodents Implicates a Multi-host Reservoir of Leishmania (Viannia) braziliensis. PLoS Negl Trop Dis 9(10): e0004137. doi:10.1371/journal. pntd.0004137 Editor: Shaden Kamhawi, National Institutes of Health, UNITED STATES Received: March 31, 2015
Abstract Background The possibility that a multi-host wildlife reservoir is responsible for maintaining transmission of Leishmania (Viannia) braziliensis causing human cutaneous and mucocutaneous leishmaniasis is tested by comparative analysis of infection progression and infectiousness to sandflies in rodent host species previously shown to have high natural infection prevalences in both sylvatic or/and peridomestic habitats in close proximity to humans in northeast Brazil.
Accepted: September 10, 2015 Published: October 8, 2015
Methods
Copyright: © 2015 Andrade et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
The clinical and parasitological outcomes, and infectiousness to sandflies, were observed in 54 colonized animals of three species (18 Necromys lasiurus, 18 Nectomys squamipes and 18 Rattus rattus) experimentally infected with high (5.5×106/ml) or low (2.8×105/ml) dose L. (V.) braziliensis (MBOL/BR/2000/CPqAM95) inoculum. Clinical signs of infection were monitored daily. Whole animal xenodiagnoses were performed 6 months post inoculation using Lutzomyia longipalpis originating from flies caught in Passira, Pernambuco, after this parasite evaluation was performed at necropsy. Heterogeneities in Leishmania parasite loads were measured by quantitative PCR in ear skin, liver and spleen tissues.
Data Availability Statement: All relevant data are within the paper. Funding: This study received financial support of Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) agency, grants 479046/2011-5 and 400446/2012-0. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist.
Results All three rodent species proved to establish infection characterized by short-term selfresolving skin lesions, located on ears and tail but not on footpads (one site of inoculation), and variable parasite loads detected in all three tissues with maximum burdens of 8.1×103 (skin), 2.8×103 (spleen), and 8.9×102 (liver). All three host species, 18/18 N. lasiurus, 10/18
PLOS Neglected Tropical Diseases | DOI:10.1371/journal.pntd.0004137
October 8, 2015
1 / 14
Infectiousness of Small Rodents to Leishmania braziliensis
N. squamipes and 6/18 R. rattus, also proved infectious to sandflies in cross-sectional study. R. rattus supported significantly lower tissue parasite loads compared to those in N. lasiurus and N. squamipes, and N. lasiurus appeared to be more infectious, on average, than either N. squamipes or R. rattus.
Conclusions A multi-host reservoir of cutaneous leishmaniasis is indicated in this region of Brazil, though with apparent differences in the competence between the rodent species. The results provide preliminary insights into links between sylvatic and peri-domestic transmission cycles associated with overlaps in the rodent species’ ecological niches.
Author Summary Across the Americas, Leishmania (V.) braziliensis is the predominant Leishmania species causing cutaneous and mucocutaneous leishmaniasis in humans. Transmitted by Phlebotomine sandflies, questions remain about the epidemiological contributions of the numerous zoonotic and more domestic host species. Domestication of the principal vector and human infection patterns suggest that human infection risk is predominantly peridomestic, whereas control strategies will be more complex if there is a link to a wildlife transmission cycle. Almost no studies have been conducted on the transmission potential of natural hosts of L. (V.) braziliensis. This study evaluates the infectiousness of experimentally infected natural rodent host species, that in different ecological habitats are proposed to act as a single or a multi-host reservoir. Clinical and parasitological development, and the ability to transmit Leishmania to sandflies, was observed under experimental conditions using a single strain of L. (V.) braziliensis isolated from the wild rat, Necromys lasiurus. Xenodiagnoses were performed with laboratory bred sand fly females established from a local population of Lutzomyia longipalpis. All three rodent species developed disseminated subclinical parasitological infections, but clinical signs (lesions) were transient and self-resolving. N. squamipes, N. lasiurus and R. rattus were all infectious when asymptomatic, though their competence in transmission potential appears to differ with R. rattus showing signs of lower susceptibility. These results provide further evidence that a multihost reservoir is responsible for maintaining transmission with a bridge between infectious sylvatic and peridomestic rodent populations.
Introduction Transmission of zoonotic pathogens may involve one, or typically more than one, reservoir host. Compared to pathogens with single reservoir hosts, those involving multi-host communities usually show reduced transmission rates through a process of zooprophylaxis or “dilution effect” due to heterogeneities in their competence to support pathogen replication and in their infectious duration, resulting in reduced pathogen-host contact, or vector-infectious host contact in the case of vector-borne pathogens [1, 2, 3]. The less common case in nature is that multi-host communities are more homogeneous as competent reservoirs, such that transmission is amplified, otherwise known as zoopotentiation; complexities in these scenarios are discussed elsewhere [2, 4]. Quantification of host heterogeneity has led to a better understanding
PLOS Neglected Tropical Diseases | DOI:10.1371/journal.pntd.0004137
October 8, 2015
2 / 14
Infectiousness of Small Rodents to Leishmania braziliensis
of transmission dynamics [1, 5, 6], and improved mathematical predictions of transmission hotspots towards development of disease surveillance and control strategies [7, 8]. Zoonotic cutaneous leishmaniasis (ZCL) is a prime example where infection has been detected in multiple host species in different habitats, but where the competence of hosts and sand fly vectors in putative transmission cycles, are not well defined. Across the Americas, the predominant aetiological agent of ZCL is L. (Viannia.) braziliensis causing, in humans, small simple self-healing cutaneous lesions to disfiguring and destructive lesions known as espundia or mucosal leishmaniasis that can result in irreversible disfigurement of the upper nasal tract. In Brazil the dominant parasite causing cutaneous leishmaniasis is L. (V.) braziliensis and there are approximately 26,000 reported new human cases per year but estimates of annual incidences range from 72,800 to 119,600 [9]. L. (V.) braziliensis infections have been identified in sylvatic vectors and small mammals in the Atlantic rainforest biome [10, 11], however transmission has expanded into anthropogenic habitats where infection is observed in more synanthropic and peridomestic species including rodents, marsupials, domestic dogs and equids [11, 12, 13] that may or may not be epidemiologically significant for transmission to humans. Human transmission is predominantly peridomestic as indicated by case age distributions e.g. not limited to adults, forest or plantation workers [14], and the known vector Lu. whitmani, is captured in large numbers in animals sheds [11, 15]. Control of human ZCL currently relies on human case detection and treatment, however since humans are not thought to be particularly infectious, interrupting transmission necessarily relies on reservoir or/and vector control. There are no comparative transmission studies of L. (V.) braziliensis in small mammal that are indicated as being natural hosts. By experimental infection, this study aims to compare the reservoir competence of wild and synanthropic rodents previously implicated as reservoirs of L. (V.) braziliensis in northeast Brazil 25. These experiments provide the initial data towards defining their individual vs collective susceptibility to infection, ability to support parasite replication, and their infectiousness to phlebotomine sand flies for onward transmission.
Materials and Methods Three rodent species (Necromys lasiurus syn. Bolomys lasiurus (Lund, 1840), Nectomys squamipes Brants, 1827 and Rattus rattus Linnaeus, 1758) were selected for comparative study as potentially important reservoirs based on demonstrating high prevalences of natural infection or/and high population densities, in previous field studies in endemic ZCL foci in Pernambuco, northeast Brazil [10, 11]. In this region, the reported incidence is 18.5 human cases per 100,000 inhabitants [16]. N. lasiurus and N. squamipes are usually associated with Atlantic rain forest and scrub/plantation habitats, whereas R. rattus is predominantly captured inside houses and animal sheds [11]. Rodent colonies were established at Fiocruz-PE from adult animals live-captured in a well studied foci, Raiz de Dentro, in the Municipality of Amaraji, Pernambuco, northeast Brazil (8° 23’S, 35° 27’W), and identified based on morphological and morphometric characters [17]. A total of 60 F1 generation 35–45 day old animals (20 R. rattus, 20 N. lasiurus and 20 N. squamipes), were selected and divided into two groups of 10 animals per species for experimental infection with a high or low dose inoculation of L. (V.) braziliensis, as described below. A control group of two hamsters (Mesocricetus auratus) per group was included to confirm the infectivity of the inoculation cultures. The L. (V.) braziliensis strain (MBOL/BR/2000/CPqAM95) used throughout our experiments was isolated on 08/06/2000 from a N. lasiurus captured in the Amaraji region, and identified as belonging to zymodeme IOC-Z74, variant 4 and serodeme 1 [18]. This same
PLOS Neglected Tropical Diseases | DOI:10.1371/journal.pntd.0004137
October 8, 2015
3 / 14
Infectiousness of Small Rodents to Leishmania braziliensis
zymodeme has been isolated from man in the endemic area of Amaraji It was cryopreserved after isolation and then ampoules were thawed and contents passaged from 3 to 4 times before inoculation. In culture and hamsters this strain behaves similarly to other strains of this parasite isolated from other wild rodents and man. Promastigotes were grown in biphasic medium of Blood Agar Base (Difco 45) [19] and Schneider's medium enriched with 20% fetal bovine serum and maintained at 26°C. For experimental infections, the inoculum was prepared from a 7 day old log phase culture containing 2.8×105/ml (low dose) or 5.5×106/ml (high dose) final concentrations according to the doses used for infecting mice with different Leishmania species[20]. 7 day old cultures are the inter phase between the log and stationary phases and are composed of infectious metacyclic promastigotes and non-infectious procyclic promastigotes. Twenty replicate animals of each species, N. lasiurus, N. squamipes and R. rattus, were inoculated experimentally with either high (n = 10) or low dose (n = 10) L. (V.) braziliensis Each animal was inoculated in the following sites: left hind paw (0.025 ml), left ear (0.025 mL), and intraperitoneal space (0.05 mL) following a protocol used for experimentally infecting Proechimys[21]. Control groups of hamsters (Mesocricetus auratus) were inoculated under the same conditions. Animals were followed up for 180 days post inoculation when submitted to xenodiagnoses, and then sacrificed as described below.
Clinical analysis After experimental infection, animals were monitored daily to detect any clinical changes including lesions on the inoculation site, hair loss, or splenomegaly.
Xenodiagnosis Xenodiagnosis was performed on 18 of each rodent species six months after inoculation using 7-day old sand flies, from the first generation Lu. longipalpis captured in a well studied foci, in the Municipality of Passira, Pernambuco, northeast Brazil (7° 56’S, 35° 35’W). The mating song of this population has been determined as a burst type, being very similar to Camara and Bacarena populations of Pará State [22]. Burst song populations are principally coastal and all have the cembrene-1 pheromone[23]. The animals were anesthetized with ketamine hydrochloride at 10% and placed in cages into which female sand flies were released and allowed to feed for about 40 minutes in the presence of a similar number of male sand flies in order to induce feeding and copulation. Blood-fed females were then transferred to plastic pots that were stored in boxes with light filter protection and kept under controlled laboratory conditions until the seventh day when they were dissected to detect promastigote forms under optical microscopy.
Necropsy and parasitology procedures After conclusion of xenodiagnosis, the animals were euthanized with a CO2 inhalation process. Fragments of approximately 50mg of ear skin, spleen and liver were collected from each euthanized animals, and Leishmania parasite DNA quantified by quantitative PCR (qPCR).
DNA extraction and molecular detection of parasites by qPCR DNA was extracted from tissues using DNeasy Blood & Tissue kit (Qiagen) according to the manufacturer’s protocol. The initial molecular detection protocol consisted of a nested PCR assay using two pairs of SSU rDNA (Small Subunit Ribosomal gene) derived oligonucleotides. The first PCR used SSU rDNA primers [24] that amplify a conserved region of all trypanosomatids (S12: 5’-GGTTGATTCCGTCAACGGAC-3’ and S4: 5’-GATCCAGCTGCAGGTTCA CC-3’); internal oligonucleotides PCR products were analyzed by electrophoresis in agarose
PLOS Neglected Tropical Diseases | DOI:10.1371/journal.pntd.0004137
October 8, 2015
4 / 14
Infectiousness of Small Rodents to Leishmania braziliensis
gel. The second reaction was a real time PCR (qPCR) to quantify the parasite load [25] using primers that amplify a common region of the Leishmania (Viannia) subgenus (S17: 5’-CCAAGC TGCCCAGTAGAAT-3’ and S18: 5’-TCGGGCGGATAAAACACC-3’). The quantification protocol consisted of a real time SYBR-Green PCR; tissue parasite loads were standardized as number of SSU rDNA copies per host glyceraldehyde-3-phosphate dehydrogenase (GAPDH) copy number. The PCR conditions were optimized to generate a single melting curve of the product.
Data analysis Established experimental infection was defined as the presence of one or more condition: development of skin lesions associated with symptomatic rodent ZCL, detection of splenomegaly at necropsy, qPCR detection of Leishmania in tissue samples (ear skin, liver, spleen), or infectiousness to sand flies. For statistical analyses, Leishmania loads were log10+1 transformed and tested using general linearised Poisson models (negative binomial over-dispersion coefficient α2.88, P< = 0.004), but not dissimilar between N. squamipes and R. rattus (z = 0.85, P = 0.398). All lesions spontaneously recovered within one month of onset, after an average 14 (95% CI: 10.9–17.1), 21 (17.8–23.5) and 19 (6.3–31.7) days for the three species, respectively. Lesion duration was shorter (i.e. faster recovery) in N. lasiurus compared to N. squamipes or to R. rattus (z>2.01, P< = 0.044), but not statistically different between N. squamipes and R. rattus (z = -0.58, P = 0.561). Table 2. Crude proportions of blood-fed dissected female sand flies that were infected by rodents experimentally inoculated with high or low dose L. braziliensis. Experimental inoculation dose
N animals xenopositive/tested (n blood-fed sand flies dissected)
Median proportion of flies infected per rodent binomial 95% C.I.s
High dose Necromys lasiurus
8/8 (77)
0.42 0.115–0.538
Nectomys squamipes
1/8 (60)
0.00 0.00–0.036
Rattus rattus
0/9 (162)
0.00 0.00–0.00
Low dose Necromys lasiurus
10/10 (206)
0.36 0.240–0.443
Nectomys squamipes
9/10 (141)
0.35 0.129–0.406
Rattus rattus
6/9 (101)
0.29 0.00–0.416
doi:10.1371/journal.pntd.0004137.t002
PLOS Neglected Tropical Diseases | DOI:10.1371/journal.pntd.0004137
October 8, 2015
6 / 14
Infectiousness of Small Rodents to Leishmania braziliensis
Fig 1. Bar graph showing the duration of skin lesions in days of laboratory bred Necromys lasiurus, Nectomys squamipes and Rattus rattus inoculated with different concentrations (High dose = 5.5×106/ml and Low dose = 2.8×105/ml) of Leishmania (Viannia) braziliensis (strain (MBOL/ BR/2000/CPqAM95). doi:10.1371/journal.pntd.0004137.g001
Parasitology The results of tissue parasite loads were quantified by qPCR in single skin, spleen, and liver tissue samples from all follow-up animals at necropsy are show in Table 3. Maximum tissue burdens were 8.1×103 in skin, 2.8×103 in spleen, and 8.9×102 in liver samples. Substantial variation in Leishmania loads were observed between individual tissues, animals, and inocula dose (Table 3 and Fig 2). Log10 parasite loads in the three tissues were only moderately correlated (Spearman’s r = 0.64–0.67, P-4.1, P-4.9, P
Infectiousness of Sylvatic and Synanthropic Small Rodents Implicates a Multi-host Reservoir of Leishmania (Viannia) braziliensis Maria S. Andrade1,2☯, Orin Courtenay3☯, Maria E. F. Brito1, Francisco G. Carvalho1, Ana Waléria S. Carvalho1, Fábia Soares1, Silvia M. Carvalho1, Pietra L. Costa1, Ricardo Zampieri4, Lucile M. Floeter-Winter4, Jeffrey J. Shaw5‡*, Sinval P. BrandãoFilho1‡* 1 Department of Immunology, Aggeu Magalhães Research Center-Oswaldo Cruz Foundation (FIOCRUZ), Recife, Pernambuco, Brazil, 2 University of Pernambuco, Recife, Pernambuco, Brazil, 3 Warwick Infectious Disease Epidemiology Research (WIDER) and School of Life Sciences, University Warwick, Coventry, United Kingdom, 4 Biosiences Institute, São Paulo University, São Paulo, São Paulo, Brazil, 5 Biomedical Institute, São Paulo University, São Paulo, São Paulo, Brazil ☯ These authors contributed equally to this work. ‡ JJS and SPBF also contributed equally to this work. * [email protected] (JJS); [email protected] SPBF OPEN ACCESS Citation: Andrade MS, Courtenay O, F. Brito ME, Carvalho FG, Carvalho AWS, Soares F, et al. (2015) Infectiousness of Sylvatic and Synanthropic Small Rodents Implicates a Multi-host Reservoir of Leishmania (Viannia) braziliensis. PLoS Negl Trop Dis 9(10): e0004137. doi:10.1371/journal. pntd.0004137 Editor: Shaden Kamhawi, National Institutes of Health, UNITED STATES Received: March 31, 2015
Abstract Background The possibility that a multi-host wildlife reservoir is responsible for maintaining transmission of Leishmania (Viannia) braziliensis causing human cutaneous and mucocutaneous leishmaniasis is tested by comparative analysis of infection progression and infectiousness to sandflies in rodent host species previously shown to have high natural infection prevalences in both sylvatic or/and peridomestic habitats in close proximity to humans in northeast Brazil.
Accepted: September 10, 2015 Published: October 8, 2015
Methods
Copyright: © 2015 Andrade et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
The clinical and parasitological outcomes, and infectiousness to sandflies, were observed in 54 colonized animals of three species (18 Necromys lasiurus, 18 Nectomys squamipes and 18 Rattus rattus) experimentally infected with high (5.5×106/ml) or low (2.8×105/ml) dose L. (V.) braziliensis (MBOL/BR/2000/CPqAM95) inoculum. Clinical signs of infection were monitored daily. Whole animal xenodiagnoses were performed 6 months post inoculation using Lutzomyia longipalpis originating from flies caught in Passira, Pernambuco, after this parasite evaluation was performed at necropsy. Heterogeneities in Leishmania parasite loads were measured by quantitative PCR in ear skin, liver and spleen tissues.
Data Availability Statement: All relevant data are within the paper. Funding: This study received financial support of Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) agency, grants 479046/2011-5 and 400446/2012-0. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist.
Results All three rodent species proved to establish infection characterized by short-term selfresolving skin lesions, located on ears and tail but not on footpads (one site of inoculation), and variable parasite loads detected in all three tissues with maximum burdens of 8.1×103 (skin), 2.8×103 (spleen), and 8.9×102 (liver). All three host species, 18/18 N. lasiurus, 10/18
PLOS Neglected Tropical Diseases | DOI:10.1371/journal.pntd.0004137
October 8, 2015
1 / 14
Infectiousness of Small Rodents to Leishmania braziliensis
N. squamipes and 6/18 R. rattus, also proved infectious to sandflies in cross-sectional study. R. rattus supported significantly lower tissue parasite loads compared to those in N. lasiurus and N. squamipes, and N. lasiurus appeared to be more infectious, on average, than either N. squamipes or R. rattus.
Conclusions A multi-host reservoir of cutaneous leishmaniasis is indicated in this region of Brazil, though with apparent differences in the competence between the rodent species. The results provide preliminary insights into links between sylvatic and peri-domestic transmission cycles associated with overlaps in the rodent species’ ecological niches.
Author Summary Across the Americas, Leishmania (V.) braziliensis is the predominant Leishmania species causing cutaneous and mucocutaneous leishmaniasis in humans. Transmitted by Phlebotomine sandflies, questions remain about the epidemiological contributions of the numerous zoonotic and more domestic host species. Domestication of the principal vector and human infection patterns suggest that human infection risk is predominantly peridomestic, whereas control strategies will be more complex if there is a link to a wildlife transmission cycle. Almost no studies have been conducted on the transmission potential of natural hosts of L. (V.) braziliensis. This study evaluates the infectiousness of experimentally infected natural rodent host species, that in different ecological habitats are proposed to act as a single or a multi-host reservoir. Clinical and parasitological development, and the ability to transmit Leishmania to sandflies, was observed under experimental conditions using a single strain of L. (V.) braziliensis isolated from the wild rat, Necromys lasiurus. Xenodiagnoses were performed with laboratory bred sand fly females established from a local population of Lutzomyia longipalpis. All three rodent species developed disseminated subclinical parasitological infections, but clinical signs (lesions) were transient and self-resolving. N. squamipes, N. lasiurus and R. rattus were all infectious when asymptomatic, though their competence in transmission potential appears to differ with R. rattus showing signs of lower susceptibility. These results provide further evidence that a multihost reservoir is responsible for maintaining transmission with a bridge between infectious sylvatic and peridomestic rodent populations.
Introduction Transmission of zoonotic pathogens may involve one, or typically more than one, reservoir host. Compared to pathogens with single reservoir hosts, those involving multi-host communities usually show reduced transmission rates through a process of zooprophylaxis or “dilution effect” due to heterogeneities in their competence to support pathogen replication and in their infectious duration, resulting in reduced pathogen-host contact, or vector-infectious host contact in the case of vector-borne pathogens [1, 2, 3]. The less common case in nature is that multi-host communities are more homogeneous as competent reservoirs, such that transmission is amplified, otherwise known as zoopotentiation; complexities in these scenarios are discussed elsewhere [2, 4]. Quantification of host heterogeneity has led to a better understanding
PLOS Neglected Tropical Diseases | DOI:10.1371/journal.pntd.0004137
October 8, 2015
2 / 14
Infectiousness of Small Rodents to Leishmania braziliensis
of transmission dynamics [1, 5, 6], and improved mathematical predictions of transmission hotspots towards development of disease surveillance and control strategies [7, 8]. Zoonotic cutaneous leishmaniasis (ZCL) is a prime example where infection has been detected in multiple host species in different habitats, but where the competence of hosts and sand fly vectors in putative transmission cycles, are not well defined. Across the Americas, the predominant aetiological agent of ZCL is L. (Viannia.) braziliensis causing, in humans, small simple self-healing cutaneous lesions to disfiguring and destructive lesions known as espundia or mucosal leishmaniasis that can result in irreversible disfigurement of the upper nasal tract. In Brazil the dominant parasite causing cutaneous leishmaniasis is L. (V.) braziliensis and there are approximately 26,000 reported new human cases per year but estimates of annual incidences range from 72,800 to 119,600 [9]. L. (V.) braziliensis infections have been identified in sylvatic vectors and small mammals in the Atlantic rainforest biome [10, 11], however transmission has expanded into anthropogenic habitats where infection is observed in more synanthropic and peridomestic species including rodents, marsupials, domestic dogs and equids [11, 12, 13] that may or may not be epidemiologically significant for transmission to humans. Human transmission is predominantly peridomestic as indicated by case age distributions e.g. not limited to adults, forest or plantation workers [14], and the known vector Lu. whitmani, is captured in large numbers in animals sheds [11, 15]. Control of human ZCL currently relies on human case detection and treatment, however since humans are not thought to be particularly infectious, interrupting transmission necessarily relies on reservoir or/and vector control. There are no comparative transmission studies of L. (V.) braziliensis in small mammal that are indicated as being natural hosts. By experimental infection, this study aims to compare the reservoir competence of wild and synanthropic rodents previously implicated as reservoirs of L. (V.) braziliensis in northeast Brazil 25. These experiments provide the initial data towards defining their individual vs collective susceptibility to infection, ability to support parasite replication, and their infectiousness to phlebotomine sand flies for onward transmission.
Materials and Methods Three rodent species (Necromys lasiurus syn. Bolomys lasiurus (Lund, 1840), Nectomys squamipes Brants, 1827 and Rattus rattus Linnaeus, 1758) were selected for comparative study as potentially important reservoirs based on demonstrating high prevalences of natural infection or/and high population densities, in previous field studies in endemic ZCL foci in Pernambuco, northeast Brazil [10, 11]. In this region, the reported incidence is 18.5 human cases per 100,000 inhabitants [16]. N. lasiurus and N. squamipes are usually associated with Atlantic rain forest and scrub/plantation habitats, whereas R. rattus is predominantly captured inside houses and animal sheds [11]. Rodent colonies were established at Fiocruz-PE from adult animals live-captured in a well studied foci, Raiz de Dentro, in the Municipality of Amaraji, Pernambuco, northeast Brazil (8° 23’S, 35° 27’W), and identified based on morphological and morphometric characters [17]. A total of 60 F1 generation 35–45 day old animals (20 R. rattus, 20 N. lasiurus and 20 N. squamipes), were selected and divided into two groups of 10 animals per species for experimental infection with a high or low dose inoculation of L. (V.) braziliensis, as described below. A control group of two hamsters (Mesocricetus auratus) per group was included to confirm the infectivity of the inoculation cultures. The L. (V.) braziliensis strain (MBOL/BR/2000/CPqAM95) used throughout our experiments was isolated on 08/06/2000 from a N. lasiurus captured in the Amaraji region, and identified as belonging to zymodeme IOC-Z74, variant 4 and serodeme 1 [18]. This same
PLOS Neglected Tropical Diseases | DOI:10.1371/journal.pntd.0004137
October 8, 2015
3 / 14
Infectiousness of Small Rodents to Leishmania braziliensis
zymodeme has been isolated from man in the endemic area of Amaraji It was cryopreserved after isolation and then ampoules were thawed and contents passaged from 3 to 4 times before inoculation. In culture and hamsters this strain behaves similarly to other strains of this parasite isolated from other wild rodents and man. Promastigotes were grown in biphasic medium of Blood Agar Base (Difco 45) [19] and Schneider's medium enriched with 20% fetal bovine serum and maintained at 26°C. For experimental infections, the inoculum was prepared from a 7 day old log phase culture containing 2.8×105/ml (low dose) or 5.5×106/ml (high dose) final concentrations according to the doses used for infecting mice with different Leishmania species[20]. 7 day old cultures are the inter phase between the log and stationary phases and are composed of infectious metacyclic promastigotes and non-infectious procyclic promastigotes. Twenty replicate animals of each species, N. lasiurus, N. squamipes and R. rattus, were inoculated experimentally with either high (n = 10) or low dose (n = 10) L. (V.) braziliensis Each animal was inoculated in the following sites: left hind paw (0.025 ml), left ear (0.025 mL), and intraperitoneal space (0.05 mL) following a protocol used for experimentally infecting Proechimys[21]. Control groups of hamsters (Mesocricetus auratus) were inoculated under the same conditions. Animals were followed up for 180 days post inoculation when submitted to xenodiagnoses, and then sacrificed as described below.
Clinical analysis After experimental infection, animals were monitored daily to detect any clinical changes including lesions on the inoculation site, hair loss, or splenomegaly.
Xenodiagnosis Xenodiagnosis was performed on 18 of each rodent species six months after inoculation using 7-day old sand flies, from the first generation Lu. longipalpis captured in a well studied foci, in the Municipality of Passira, Pernambuco, northeast Brazil (7° 56’S, 35° 35’W). The mating song of this population has been determined as a burst type, being very similar to Camara and Bacarena populations of Pará State [22]. Burst song populations are principally coastal and all have the cembrene-1 pheromone[23]. The animals were anesthetized with ketamine hydrochloride at 10% and placed in cages into which female sand flies were released and allowed to feed for about 40 minutes in the presence of a similar number of male sand flies in order to induce feeding and copulation. Blood-fed females were then transferred to plastic pots that were stored in boxes with light filter protection and kept under controlled laboratory conditions until the seventh day when they were dissected to detect promastigote forms under optical microscopy.
Necropsy and parasitology procedures After conclusion of xenodiagnosis, the animals were euthanized with a CO2 inhalation process. Fragments of approximately 50mg of ear skin, spleen and liver were collected from each euthanized animals, and Leishmania parasite DNA quantified by quantitative PCR (qPCR).
DNA extraction and molecular detection of parasites by qPCR DNA was extracted from tissues using DNeasy Blood & Tissue kit (Qiagen) according to the manufacturer’s protocol. The initial molecular detection protocol consisted of a nested PCR assay using two pairs of SSU rDNA (Small Subunit Ribosomal gene) derived oligonucleotides. The first PCR used SSU rDNA primers [24] that amplify a conserved region of all trypanosomatids (S12: 5’-GGTTGATTCCGTCAACGGAC-3’ and S4: 5’-GATCCAGCTGCAGGTTCA CC-3’); internal oligonucleotides PCR products were analyzed by electrophoresis in agarose
PLOS Neglected Tropical Diseases | DOI:10.1371/journal.pntd.0004137
October 8, 2015
4 / 14
Infectiousness of Small Rodents to Leishmania braziliensis
gel. The second reaction was a real time PCR (qPCR) to quantify the parasite load [25] using primers that amplify a common region of the Leishmania (Viannia) subgenus (S17: 5’-CCAAGC TGCCCAGTAGAAT-3’ and S18: 5’-TCGGGCGGATAAAACACC-3’). The quantification protocol consisted of a real time SYBR-Green PCR; tissue parasite loads were standardized as number of SSU rDNA copies per host glyceraldehyde-3-phosphate dehydrogenase (GAPDH) copy number. The PCR conditions were optimized to generate a single melting curve of the product.
Data analysis Established experimental infection was defined as the presence of one or more condition: development of skin lesions associated with symptomatic rodent ZCL, detection of splenomegaly at necropsy, qPCR detection of Leishmania in tissue samples (ear skin, liver, spleen), or infectiousness to sand flies. For statistical analyses, Leishmania loads were log10+1 transformed and tested using general linearised Poisson models (negative binomial over-dispersion coefficient α2.88, P< = 0.004), but not dissimilar between N. squamipes and R. rattus (z = 0.85, P = 0.398). All lesions spontaneously recovered within one month of onset, after an average 14 (95% CI: 10.9–17.1), 21 (17.8–23.5) and 19 (6.3–31.7) days for the three species, respectively. Lesion duration was shorter (i.e. faster recovery) in N. lasiurus compared to N. squamipes or to R. rattus (z>2.01, P< = 0.044), but not statistically different between N. squamipes and R. rattus (z = -0.58, P = 0.561). Table 2. Crude proportions of blood-fed dissected female sand flies that were infected by rodents experimentally inoculated with high or low dose L. braziliensis. Experimental inoculation dose
N animals xenopositive/tested (n blood-fed sand flies dissected)
Median proportion of flies infected per rodent binomial 95% C.I.s
High dose Necromys lasiurus
8/8 (77)
0.42 0.115–0.538
Nectomys squamipes
1/8 (60)
0.00 0.00–0.036
Rattus rattus
0/9 (162)
0.00 0.00–0.00
Low dose Necromys lasiurus
10/10 (206)
0.36 0.240–0.443
Nectomys squamipes
9/10 (141)
0.35 0.129–0.406
Rattus rattus
6/9 (101)
0.29 0.00–0.416
doi:10.1371/journal.pntd.0004137.t002
PLOS Neglected Tropical Diseases | DOI:10.1371/journal.pntd.0004137
October 8, 2015
6 / 14
Infectiousness of Small Rodents to Leishmania braziliensis
Fig 1. Bar graph showing the duration of skin lesions in days of laboratory bred Necromys lasiurus, Nectomys squamipes and Rattus rattus inoculated with different concentrations (High dose = 5.5×106/ml and Low dose = 2.8×105/ml) of Leishmania (Viannia) braziliensis (strain (MBOL/ BR/2000/CPqAM95). doi:10.1371/journal.pntd.0004137.g001
Parasitology The results of tissue parasite loads were quantified by qPCR in single skin, spleen, and liver tissue samples from all follow-up animals at necropsy are show in Table 3. Maximum tissue burdens were 8.1×103 in skin, 2.8×103 in spleen, and 8.9×102 in liver samples. Substantial variation in Leishmania loads were observed between individual tissues, animals, and inocula dose (Table 3 and Fig 2). Log10 parasite loads in the three tissues were only moderately correlated (Spearman’s r = 0.64–0.67, P-4.1, P-4.9, P
